DE102006046206A1 - A connection device for randomly connecting a number of transmitters and receivers, communication device and method for establishing a connection device - Google Patents

Abstract

A connection device for random connection of a first number of first transceiver units (SE1-1, SE2-1, SE3-1, SE4-1) to a second number of second transceiver units (SE1-2, SE2-) is provided. 2, SE3-2, SE4-2). The connection device comprises a switch matrix (SM) with a third number of controllable mh control of a switching element (MMS), an electrical connection between one of the first transceiver units (SE1-1, SE2-1, SE3-1, SE4-1) and one of the second transmitting / receiving units (SE1-2, SE2-2, SE3-2, SE4-2) can be produced, and a drive circuit (AS) for selectively controlling each of the micromechanical switching elements (MMS).

Description

The The invention relates to a connection device for random connection a first number of first transceiver units having a second one Number of second transceiver units. The invention further relates to a communication device with at least a first and at least one second transceiver unit and with a connection device for selective production an electrical connection between the at least one first and the at least one second transceiver unit. The invention relates Furthermore, a method for producing a connecting device.

communication facilities for transmission of information comprises at least a first transceiver unit and at least one second transceiver unit. In the simplest Case is in each case one of the first transceiver units with a the second transceiver units connected to a channel. at Communication devices with multiple transmission channels is regarding the optimization of costs and the necessary transmission lines a multiplexing advantageous. In this case, several first transmitting / receiving units optionally switched to a second transceiver unit, which reduces the number of system components. In other Situations it is necessary to share information between a plurality at first transmitting / receiving units and a plurality at second Transmitter / receiver units to transmit in parallel wherein the assignment of first transmitting / receiving units and second Transmitter / receiver units must be optional. This applies, for example, in a telephone exchange. Such a requirement leaves comply with a space division multiplexing.

1 shows a schematic representation of a so-called. Cross bar distribution in a communication device. In 1 First transceiver units are identified by the reference symbols SE1-1, SE2-1, SE3-1 and SE4-1. The first transceivers SE1-1, ..., SE4-1 are coupled to input lines LE1, LE2, LE3, LE4. Second transmitting / receiving units are identified by the reference symbols SE1-2, SE2-2, SE3-2 and SE4-2. The second transceiver units SE1-2, ..., SE4-2 are each coupled to an output line LA1, LA2, LA3, LA4. In this embodiment, the number of first transmitting / receiving units SE1-1, ..., SE4-1 is selected according to the number of second transmitting / receiving units SE1-2, ..., SE4-2. This is just an example. The number of first and second receiving units can in principle be chosen arbitrarily.

to optional connection of the first transceiver units SE1-1, ..., SE4-1 and the second transceiver units SE1-2, ..., SE4-2, a switch matrix SM is provided. The switch matrix SM comprises at the nodes K11, K12, K13, K14, ..., K41, K42, K43, K44 (in general Kij, where i is the number of input lines and j is the number of output lines) each a switching element (not shown). By a arranged at a node Kij Switching element is doing an electrical connection in the made the node crossing input line and output line.

By appropriate control of respective switching elements by a in the 1 Not shown drive circuit can each be parallel to a certain one of the first transceiver units SE1-1, ..., SE4-1 electrically connect to a certain one of the second transceiver units SE1-2, ..., SE4-2. This is in 1 represented by the designated by the reference numerals P1, ..., P4 communication paths by way of example. With the communication path P1, the first transmitting / receiving unit SE1-1 is connected to the second transmitting / receiving unit SE2-2. In this case, the first transmitting / receiving unit SE1-1 is a transmitter and the second transmitting / receiving unit SE2-2 is a receiver, which is expressed by the arrow direction. The electrical connection takes place in that the switching element arranged at the node K12 is turned on, while the other switching elements arranged in the communication path T1 are held in a blocking manner in the nodes K11, K22, K32 and K42.

In Similarly, a communication path P2 between the first transmitting / receiving unit SE2-1 and the second transmitting / receiving unit SE1-2, between the first transmitting / receiving unit SE3-1 and the second transmitting / receiving unit SE3-2 and the first transmitting / receiving unit SE4-1 and the second transmitting / receiving unit SE4-2. In the illustrated embodiment set the first transceiver units SE2-1 and SE4-1 recipients while SE3-1 forms a transmitter. Correspondingly, the second Transceiver units SE1-2 and SE4-2 transmitters and SE3-2 one receiver represents.

A general requirement for one, as in 1 described switch matrix is that the multiplexing process does not affect the actual signal transmission. The switching elements necessary for the multiplexing have parasitic elements, a resistance and a capacitance. In particular, the capacity for high frequencies in the range of more than 100 MHz represents an undesirable leakage path, which leads to signal losses. This results especially in the transfer radio frequency signals is a signal integrity problem.

Around the unwanted ones Signal losses to avoid, so far switching elements with low parasitic capacities used or the effect of parasitic capacitances becomes active through appropriate Circuit arrangements compensated. Switches with low parasitic capacitances are For example, PIN diodes, but the disadvantage of very high production costs exhibit. The provision of components for reducing the parasitic capacitance is while cheaper, However, has the disadvantage that the space requirement for the switch matrix increases. This takes up the space required, especially when switch matrixes for random connection, a large number of first and second Transmit / receive units have a no longer tolerable Size one.

It is therefore an object of the present invention, a connection device for the optional connection of transmitting / receiving units and a Specify communication device in which a signal transmission, in particular of high-frequency signals, with no or only small signal losses leads. At the same time they should be realized with a small footprint be.

A Another object of the invention is a method for manufacturing to specify a connecting device of the above type, which on simple and inexpensive Way is feasible.

These Tasks are solved with the features of the independent claims. advantageous embodiments each result from the dependent Claims.

A Connecting device according to the invention for random connection of a first number of transceiver units with a second number of transceiver units comprises a switch matrix, a third number of controllable micromechanical switching elements includes, by driving a switching element, an electrical Connection between one of the first transceiver units and one of the second transceiver units can be produced, and a drive circuit for selectively driving each of the micromechanical Switching elements.

Of the Reduction of signal losses at high-frequency signals at the same time minimal space requirement for a connection device according to the invention results from the use of micromechanical switching elements. These are called MEMS (Micro-Electro-Mechanical System) known and make a combination of mechanical elements and electronic elements on a substrate or semiconductor chip. Micromechanical switching elements are components with a low level parasitic capacity and thus especially for the transfer suitable for high-frequency signals. The possibilities of microsystem technology allow a miniaturized and cost-effective implementation of a switch matrix.

in the The term of the present invention is the term of the switching element to understand that includes both switches and relays are. The switching elements are each at a crossing point of Arranged and connect input lines and output lines, with appropriate control, the respective input with the Output line.

The measured third number of controllable micromechanical switching elements according to the first number of the first transceiver units, the connected to input lines and the second number to second Transceiver units connected to output lines. In general, the third number of switching elements results from the Multiplication of the first number of first transceiver units with the second number of second transceiver units. A Connecting device according to the invention thus comprises at least one switching element, if only one first transmitting / receiving unit and a second transmitting / receiving unit are provided.

The Drive circuit, for example, by an ASIC (Application Specific Integrated Circuit), which is the electrical Control of the individual switching elements of the switch matrix is used.

According to one advantageous embodiment the switch matrix and the drive circuit as integrated components in formed a common substrate. By the direct integration of the Switching elements with the components of the drive circuit results an integrated system that is inexpensive to produce and in terms his space requirement optimized for is the multiplexing of high frequency signals. The switching elements the switch matrix are expediently formed or produced in a CMOS-compatible process. The connection device is at a frequency of more than 100 MHz, more preferably operable at a frequency of more than 300 MHz.

In order to control the micromechanical switching elements, the drive circuit according to a further embodiment comprises a level matching circuit which requires the necessary for driving the micromechanical switching elements Voltage generated. Advantageously, the drive circuit comprises a position register, in which a circuit pattern for driving all the switching elements can be written. The bit width of the position register is dimensioned according to the size of the switch matrix in order to allow a simultaneous control of all nodes.

in this connection it is useful if the drive circuit comprises a memory in which a number on predetermined circuit patterns for driving the switching elements the switch matrix is deposited, wherein the stored in the memory circuit pattern the position register can be fed. The provision of the memory with the circuit patterns held therein has the advantage that a very fast control of the switch matrix possible is. Independently Of course it is conceivable that the position register via a data line with a Circuit pattern is described case by case.

in principle can the technical structure of the micromechanical switching elements to be of any nature. Preference is given to micromechanical switching elements used, which can be controlled by means of an electric field effect are. Alternatively, the micromechanical Switching elements controlled by a piezoelectric effect be.

A Communication device according to the invention comprises at least a first transmitting / receiving unit and at least a second transmitting / receiving unit and a connection device for selective production of an electrical connection between the at least one first and the at least one second transceiver unit, wherein the connection means is formed as described above is. This is accompanied by the same advantages as above have been described.

One method according to the invention for producing a connection device for the optional connection a first number of first transceiver units having a second number of second transceiver units, wherein the connection means a switch matrix comprising a third number of controllable micromechanical switching elements and a drive circuit for selectively driving each the micromechanical switching elements comprises the following Steps: Providing a first wafer in which the drive circuit is formed, and forming a first electrically conductive layer an upper surface of the first wafer, wherein the first conductive layer for each Switching element at least a first electrode for electrostatic actuation, a first electrode for a load circuit and a first electrode for connection to a drive circuit includes; Providing a second wafer and training a second electrically conductive Layer on a front side of the second wafer, wherein the second conductive Layer for each switching element at least one second electrode for electrostatic Aktuation and a second electrode for the load circuit comprises; Connect of the first and second wafers with each other such that an electric Connection between the first electrode for the connection to the drive circuit and the second electrode for the electrostatic actuation is produced.

there are the first electrode for electrostatic actuation and the first electrode for the connection connected to the drive circuit to the drive circuit. A electrical connection of the first electrode for the load circuit with the drive circuit is not necessary.

The Process for the preparation of a connecting device according to the invention can be performed using known technologies. The integration of functionality the switch matrix into a wafer, which the drive circuit is carried out by further processing of this wafer and the provision and processing of a second wafer, which in one later Processing step is connected to the first wafer. The connection takes place for example by known bonding mechanisms. The usage of manufacturing processes known from semiconductor manufacturing are possible in a simple way the formation of arbitrarily large switch matrices. The control the switching elements of the switch matrix by the drive circuit in the first wafer is over formed in the first wafer metal layers in a simple manner possible. This leaves focus on cost Way produce a connecting device, which in comparison to conventional connecting devices has a much smaller footprint.

According to one expedient embodiment the first wafer on the top before forming the first electrical conductive Layer, in particular by a chemical-mechanical method (Chemical Mechanical Polishing), planarized.

According to one further expedient embodiment is before forming the first electrically conductive layer generates at least one pit on top of the first wafer, their size is after the base area of the switching element, wherein within the at least one Pit at least one opening is formed, which expose a metal layer provided in the first wafer.

In a further method step, the first electrode for the connection to the drive circuit at least partially galvanically amplified. Via the galvanic reinforcement, the electrical connection to the electrically conductive layer on the second wafer is produced in a later method step.

Conveniently, For example, the second wafer is formed by an SOI substrate. SOI stands for silicone Insulator ("Silicon on an insulator "). The SOI technique is a known type of silicon transistor circuit. These are located on an insulating material, resulting in shorter switching times and lower power consumption. Unlike ordinary Switches that are made directly on the silicon wafer have the transistors on an insulator layer have a lower capacitance, so that which needed to switch Charges are reduced. On the so reduced switching times become higher Clock rates enabled. At the same time, the power consumption is reduced, resulting in also smaller heat loss result. For the production, e.g. a silicon wafer thermally oxidized, then becomes a second wafer with its monocrystalline surface placed on the oxide layer.

In a further process step is on the second wafer a Insulation layer applied. This will be in a subsequent Processing step structured so that the subsequently applied second electrically conductive Layer, in particular a metal layer, below the surface of the Insulation layer is arranged. The formation of the second electrodes is suitably carried out by Structuring of the second conductive Layer.

In Another method step is a boom in the second Wafer made by placing from the front of the second wafer ago a digging introduced to a buried insulation layer becomes. After all is from the back The second wafer is subjected to a grinding and etching process, so that the boom is deflected when force.

The Invention will now be described with reference to an embodiment in the drawing explained in more detail. It demonstrate:

1 a schematic representation of a crossbar distribution for performing a space division multiplex method,

2 a schematic representation of a connecting device according to the invention,

3a to d consecutive manufacturing steps in the processing of a first wafer, in which a drive circuit of the connection device is formed,

4a to d successive steps in the processing of a second wafer to form subcomponents of a micromechanical switch, and

5a and b subsequent processing steps after connecting the first and second wafers.

A connection device according to the invention 1 for performing a space division method, for example, in a crossbar distribution, is in 2 shown in a schematic way. The connection device 1 comprises a switch matrix SM and a drive circuit AS. The switch matrix SM comprises, by way of example, 8 × 8 micromechanical switching elements MMS. These are arranged at nodes of eight input lines LE1, ..., LE8 and eight output lines LA1, ..., LA8. For the sake of clarity, in the switch matrix SM the 2 only a few of the micro-mechanical switching elements MMS exemplified. Each of the switching elements MMS can be selectively controlled by the drive circuit AS.

One micromechanical switching element MMS has three electrodes, from which an electrode with the drive circuit AS, a second electrode with one of the input lines LEi, i = 1 to 8, and the third electrode are connected to an output line LAi, i = 1 to 8. At the Driving the electrode connected to the drive circuit AS a switching element MMS becomes a short circuit between the second and third electrode made so that an electrical connection between the relevant input line LEi and the relevant Output line LAi is made.

The Control of the switch matrix SM takes place in such a way that in each case only one switching element MMS of an output line LAi and an input line LEi is driven.

to Control of a switching element has the drive circuit AS a number of multiplexers MUX1, .., MUX8. In the embodiment must in each case one switching element MMS can be selected in eight switch gaps. For addressing the eight switch columns 3 bits are necessary, so that the drive over 3: 8 multiplexer takes place. For the random addressing of the 8x8 switch matrix SM thus a total of 24 bits (8 columns x 3 bits) is needed. One such "circuit pattern" is in one with the multiplexers MUX1, ..., MUX8 coupled position register PR maintained.

The position register PR can via a Da data line. The data line Data may be formed as a bus or a single line. In the exemplary embodiment, the drive circuit AS includes an internal memory SP in which a number of predefined circuit patterns are stored. The memory SP is therefore connected to the data line Data and also via a line with the position register PR. Via a selection line switch which is coupled to the output of the memory SP, it is possible to select which of the circuit patterns stored in the memory SP should be written into the position register PR. The selection line switch is 4 bits wide in the exemplary embodiment for this purpose.

The Drive circuit AS is operated with two voltage levels VDD and VSS, which at a Versorgungspotential- and a reference potential line issue. The components of the drive circuit AS work with CMOS voltage levels, usually 5V. The for controlling the micromechanical Switching elements (so-called MEMS switches or relays) needed, higher, tensions, e.g. 60V, at the output of the drive circuit AS with a level adjustment circuit LS generated. The level adjusting circuit LS is also called "level shifter" and has this Purpose over a voltage line HV at which a corresponding voltage signal is applied. About that In addition, the drive circuit AS comprises a known manner Clock generator CG, which is connected to a clock clock line is.

In the following 3 to 5 the production of a connecting device according to the invention will be described, wherein for illustrative purposes only a single switching element is shown. However, the method according to the invention is independently applicable to an entire switch matrix.

The inventive method is characterized in that the switching elements of the switch matrix together with the components of the drive circuit in a common Substrate are integrated.

Starting material (cf. 3a ) is a planarized first wafer 10 , The planarization takes place on a top 11 of the first wafer 10 , The planarization is carried out, for example, by means of chemical mechanical polishing (CMP). In the area of a back 12 of the first wafer 10 CMOS structures are formed, which playfully over the entire surface of the first wafer 10 extend. The CMOS structures are identified by the reference numeral 13 characterized. Known manner is within the substrate of the first wafer 10 formed a number of metal layers, which are at least partially interconnected via vias. By way of example and schematically, in the present embodiment, two metal layers 14 . 16 represented by vias 15 . 17 can be connected to each other.

To produce a connecting device according to the invention are in a first step from the top 11 of the first wafer 10 Two pits here 18 . 19 generated on the one hand define the base for a switching element and on the other hand, the contacting of electrodes to be generated later. Below the pits 18 . 19 are sections of the metal layers 14 . 16 arranged. Over the metal layers 14 . 16 and through contacts 15 . 17 An electrical contact of the drive circuit with the switching element to be generated takes place. For this purpose be inside the pits 18 . 19 openings 20 . 21 . 22 to the metal layer 16 generated as in 3b is shown.

This completes the step of producing an electrically conductive layer 24 on (cf. 3c ), wherein structuring the electrically conductive layer 24 a first electrode 25 for electrostatic actuation, a first electrode 26 for connection to a load circuit and a first electrode 27 be formed for the connection to the drive circuit. How out 3c is readily apparent, lies the first electrode 27 for the connection to the drive circuit in the first pit 18 while the electrodes 24 . 25 in the neighboring pit 19 are arranged. Contrary to the drawing is an electrical contact between the first electrode for the load circuit 26 and the drive circuit within the CMOS structure 13 not necessary or useful.

For the purpose of a subsequent electrical contacting is in a further process step, the in 3d is shown, a portion of the first electrode 27 galvanically amplified for the connection to the drive circuit, which is denoted by the reference numeral 28 is shown.

In 4 is the processing of a second wafer, which is formed as an SOI substrate represented. On the SOI substrate (silicon to insulator) 50 that is a first semiconductor layer 51 made of Si, an insulating layer applied thereon 52 of SiO ₂ and a further semiconductor layer applied thereto 53 Si is an insulation layer 55 applied. The insulation layer 55 For example, it may also be formed from SiO ₂ . The insulation layer 55 will be in the in 4a structured way, so that trenches 56 . 57 be formed. In the trenches 56 . 57 will, how 4b shows an electrically conductive layer 58 introduced, wherein by structuring the metal layer 58 a second electrode 59 for electrostatic actuation and a second electrode 60 created for the load circuit. As in 4c is shown on the second electrode 60 for the load circuit, a switch contact surface 61 structured by galvanic reinforcement. In a further process step, which in 4d is shown, a ditch 62 from a front side 54 of the SOI substrate 50 introduced until the insulation layer 52 is exposed. This will, as from the 5b better, a boom 63 allows.

5a shows a process step after which the wafers processed according to the previously described steps 10 and 50 connected to each other. Here are the top 11 of the first wafer 10 and the front 54 of the second wafer 50 facing each other. The connection of the first wafer 10 with the second wafer 50 is preferably carried out using a bonding method. In this case, an electrical connection of the second electrode 59 for electrostatic situation with the galvanic reinforcement 58 of the first wafer 10 produced. The second electrode 59 for electrostatic actuation is thus connected to the drive circuit.

5b shows a completed connecting device according to the invention after the second wafer 50 was ground back. Thereby, the semiconductor layer became 51 and the insulation layer 52 removed, so that now a deformable boom 63 is created. By applying appropriate potentials to the first electrode for connection to the drive circuit 27 and the first electrode 25 for electrostatic actuation will be between the electrode 25 and the portion of the second electrode 59 , the first electrode 25 opposite, generates an electric field, which leads to a force and thus deflection of the boom 63 leads. In this case, the switch contact surface passes 61 in contact with the first electrode 26 for the load circuit. Not apparent from the drawing, but readily understood by one skilled in the art, is located in front of or behind the first electrode 26 another load circuit electrode, wherein through the switch contact surface 61 an electrical short between the first electrode 26 and the unrecognizable further load circuit electrode is caused. Is the first electrode 26 For example, connected to an input line and the other load circuit electrode to an output line, so there is an electrical connection between the connected to the input line transmitting / receiving device and connected to the output line transmitting / receiving device.

Claims (20)

Connection device for random connection a first number of first transceiver units (SE1-1, SE2-1, SE3-1, SE4-1) with a second number of second transceiver units (SE1-2, SE2-2, SE3-2, SE4-2), with - one Switch matrix (SM) containing a third number of controllable micromechanical Switching elements (MMS) comprises, wherein by driving a switching element (MMS) an electrical connection between one of the first transceiver units (SE1-1, SE2-1, SE3-1, SE4-1) and one of the second transceiver units (SE1-2, SE2-2, SE3-2, SE4-2) can be produced, and - one Control circuit (AS) for selectively controlling each of the micromechanical Switching elements (MMS). Connecting device according to claim 1, wherein the Switch matrix (SM) and the drive circuit (AS) as an integrated Components are formed in a common substrate. Connecting device according to claim 1 or 2, at the switching elements (MMS) of the switch matrix (SM) in a CMOS-compatible process are formed or produced. Connecting device according to one of the previous Claims, in these with a frequency of more than 100 MHz, more preferred with a frequency of more than 300 MHz is operable. Connecting device according to one of the previous Claims, in which the drive circuit (AS) has a level matching circuit (LS) for controlling the micromechanical switching elements (MMS) includes. Connecting device according to one of the previous Claims, in which the drive circuit (AS) comprises a position register (PR), into a circuit pattern for controlling all switching elements (MMS) is writable. Connecting device according to claim 6, wherein the Drive circuit (AS) comprises a memory (SP), in which a Number of predetermined circuit patterns for driving the switching elements (MMS) of the switch matrix (SM) is stored, wherein the memory in the memory (SP) deposited circuit pattern the position register (PR) can be fed. Connecting device according to one of the previous Claims, in which the micromechanical switching elements (MMS) by means of a electric field effect can be controlled. Connecting device according to one of the previous Claims, in which the micromechanical switching elements (MMS) by means of a piezoelectric effect can be controlled. Communication device with at least one first transmitting / receiving unit (SE1-1, SE2-1, SE3-1, SE4-1) and with at least one second transmitting / receiving unit (SE1-2, SE2-2, SE3-2, SE4- 2) and with a connection device ( 1 ) for selectively establishing an electrical connection between the at least one first and the at least one second transmitting / receiving unit, in which the connecting device ( 1 ) is designed according to one of the preceding claims. Method for producing a connection device ( 1 ) for the optional connection of a first number of first transceiver units (SE1-1, SE2-1, SE3-1, SE4-1) to a second number of second transceiver units (SE1-2, SE2-2, SE3-2, SE4-2), the connecting device ( 1 ) a switch matrix (SM) comprising a third number of controllable micromechanical switching elements (MMS), and a drive circuit (AS) for selectively controlling each of the micromechanical switching elements (MMS), comprising the steps of: - providing a first wafer ( 10 ) in which the drive circuit (AS) is formed, and forming a first electrically conductive layer (FIG. 24 ) on a top side ( 11 ) of the first wafer, wherein the first conductive layer for each switching element (MMS) at least one first electrode ( 25 ) for electrostatic actuation, a first electrode ( 26 ) for a load circuit and a first electrode ( 27 ) for connection to the drive circuit (AS); Providing a second wafer ( 50 ) and forming a second electrically conductive layer ( 58 ) on a front side ( 54 ) of the second wafer, wherein the second conductive layer ( 58 ) for each switching element (MMS) at least one second electrode ( 59 ) for electrostatic actuation and a second electrode ( 60 ) for the load circuit; Connecting the first and second wafers ( 10 . 50 ) in such a way that an electrical connection between the first electrode ( 27 ) for the connection to the drive circuit (AS) and the second electrode ( 59 ) for electrostatic actuation. The method of claim 11, wherein the first wafer ( 10 ) on the top ( 11 ) before forming the first electrically conductive layer ( 24 ), in particular by a chemical-mechanical process, is planarized. Method according to claim 11 or 12, wherein prior to the formation of the first electrically conductive layer ( 24 ) on the top ( 11 ) of the first wafer ( 10 ) at least one pit ( 18 . 19 ) whose size is measured according to the base area of the switching element (MMS), wherein within the at least one pit ( 18 . 19 ) at least one opening ( 20 . 21 . 22 ), which one in the first wafer ( 11 ) provided metal layer ( 16 ) uncover. Method according to one of Claims 11 to 13, in which the first electrode ( 27 ) is at least partially galvanically amplified for the connection to the drive circuit (AS). Method according to one of Claims 11 to 14, in which the second wafer ( 50 ) is formed by an SOI substrate (SOI = silicon to insulator). Method according to one of claims 11 to 15, in which on the second wafer ( 50 ) an insulation layer ( 55 ) is applied. Method according to Claim 16, in which the insulating layer ( 55 ) is structured so that the subsequently applied second electrically conductive layer ( 58 ), in particular a metal layer, below the surface of the insulating layer ( 55 ) is arranged. Method according to one of claims 11 to 17, wherein the forming of the second electrodes ( 59 . 60 ) by structuring the second conductive layer ( 58 ) he follows. Method according to one of claims 11 to 18, wherein a boom ( 63 ) in the second wafer ( 50 ) is made by the front ( 54 ) of the second wafer ( 50 ) digging ( 62 ) to a buried isolation layer ( 52 ) is introduced. The method of claim 19, wherein the second wafer ( 50 ) is subjected from the back to a grinding and etching process, so that the boom ( 63 ) is deflected when force.